Children with autism spectrum disorder (ASD) are much more likely to wander or run away from supervision than their typically developing peers; a behavior referred to as elopement. The prevalence of elopement among children with ASD has contributed to the fact that accidents are the leading cause of premature death in this population, with twice the mortality rate of neurotypical peers. The tremendous impact elopement has on the family of a child with ASD, along with life threatening consequences for the child, have led to elopement being singled out by the Interagency Autism Coordinating Committee (IACC) as a problem in need of significant focus by researchers. There are two types of elopement: bolting and wandering. When individuals bolt they run from supervision to produce a desired outcome, such as obtaining access to preferred items or escaping a nonpreferred activity. Because restricted interests and a tendency to find certain forms of stimulation aversive are part of an ASD diagnosis it is not surprising that some children with autism are prone to this type of elopement. In contrast children who wander simply stray from supervision without a consistent motivator. Unfortunately, the treatment literature provides very little guidance or empirical support for specific treatments. Those treatment studies that do exist have all been case studies of behavioral interventions. Such treatments have shown promising results, but are also individualized and intensive, requiring significant resources and expertise t implement. Moreover, they have focused exclusively on treating bolting, whereas treatment of wandering has not been addressed in the literature to date. Thus, there is tremendous need for a comprehensive and exportable treatment for both bolting and wandering that can be implemented in a range of natural environments by trained therapists. The proposed study will build upon the clinical experience of our research group, which has treated many cases of elopement in both clinical and natural environments, to evaluate a novel intervention for elopement in children with autism: the Elopement Prevention and Safety Training (EPST) program. This intervention employs a modular format, with individual components that target the specific complexities posed by bolting or wandering. It also includes a universal component that addresses the safety concerns that are common to both types of elopement. This treatment is highly exportable and can be implemented by trained individuals with a basic level of experience in the treatment of problem behaviors in children with ASD. We assert that this modular approach to treating elopement will lead to a reduced risk of harm from elopement, and decreased parent stress related to elopement. We propose a preliminary RCT of 24 children (4-12 years) with autism to evaluate the feasibility of the proposed intervention.